Facing mounting global backlash, regulatory scrutiny and government pressure, Elon Musk-led social media platform X has disabled key image generation features of its AI chatbot Grok, following widespread misuse of the tool to create sexualised and non-consensual images of women and children.
In a statement posted by
xAI’s safety handle, @Safety, X says it has implemented new technological measures to prevent the Grok account from allowing the editing or generation of images of real people in revealing clothing, such as bikinis. The restriction applies to all users, including paid subscribers.
“We remain committed to making X a safe platform for everyone and continue to have zero tolerance for any forms of child sexual exploitation, non-consensual nudity, and unwanted sexual content,” the company says, adding that it removes high-priority violative content, including child sexual abuse material (CSAM), and reports relevant cases to law enforcement authorities where necessary.
The move follows intense criticism of Grok’s so-called 'spicy mode' which allowed users to generate sexualised deepfakes using simple text prompts such as 'put her in a bikini' or 'remove her clothes'. The feature drew sharp reactions from governments and civil society groups across several countries, with some jurisdictions blocking access to the chatbot and others launching formal probes.
On
3 January 2026, Mr Musk, in a post on X stated that anyone using Grok to make illegal content will suffer the same consequences as if they had uploaded illegal content. The next day, X issued the warning to users in similar terms, asserting that it would remove unlawful material, permanently suspend offending accounts and cooperate with local authorities and law enforcement where necessary.
In India, the ministry of electronics and information technology (MeitY) had sought an action-taken report from xAI and directed the platform to remove or disable access to unlawful content without delay. The ministry also asked X to enforce its terms of service and take strong deterrent measures, including suspending or terminating accounts misusing Grok. X reportedly removed about 3,500 obscene or sexually explicit images generated through Grok in India and barred around 600 users found misusing the tool.
Internationally, scrutiny intensified after the attorney general of California initiated an investigation into Grok’s developer, while a coalition of 28 civil society organisations wrote to Apple and Google urging them to remove Grok and X from their app stores. In the UK, prime minister Keir Starmer warned of strong regulatory action, describing the spread of AI-generated sexualised images as 'disgraceful' and 'disgusting' and backing the communications regulator Ofcom to act under the Online Safety Act, which allows fines of up to 10% of a company’s global turnover or, in extreme cases, court-ordered blocking.
As part of the latest changes, X has also restricted image creation and editing via the Grok account on the platform to paid subscribers only. According to X, this adds 'an extra layer of protection' by ensuring that users who attempt to abuse the tool can be identified and held accountable. The company also said it has introduced geoblocking to prevent users in jurisdictions where it is illegal from generating images of real people in bikinis, underwear or similar attire through Grok.
However, concerns remain that safeguards may not go far enough. Reports suggest that while the public Grok account has been heavily restricted, non-paying users may still be able to generate sexualised imagery through a separate Grok app that does not post images publicly.
Research by AI Forensics, a Paris-based non-profit organisation, found hundreds of pornographic and sexually violent images and videos created using the Grok Imagine tool, including professional-looking AI-generated videos depicting explicit sexual acts and violence.
Dr Federica Fedorczyk, an early career research fellow at the Institute for Ethics in AI, says the controversy was not the result of a sudden technical failure but reflected long-standing design choices. “From the very beginning, Grok was structurally designed to operate with fewer safeguards and guardrails than other AI assistants,” she said. According to her, concerns had been raised by victims since Grok Imagine was launched, but the issue only drew widespread attention recently after images of minors were sexualised, a bikini image of the Princess of Wales was generated and such content began circulating on X at scale.
“What we are seeing now is not an anomaly,” Dr Fedorczyk says. “These capabilities have always been present in the system, and several prior warnings were ignored. The recent wave of public outrage has merely brought long-standing problems into sharper focus.”
According to the social media platform, the latest restrictions do not change its existing safety protocols, under which all AI prompts and generated content must comply with X Rules. The company says its safety teams continue to work 'around the clock' to add safeguards, remove illegal content, suspend offending accounts and cooperate with governments and law enforcement.
As generative AI tools rapidly evolve, the Grok controversy has highlighted growing concerns about accountability, consent and the adequacy of safety-by-design measures, with regulators and experts warning that piecemeal fixes may not be enough to address systemic risks.
Chaitanya Patil’s 490Km, NH66 Audit Exposes Decades of Deadly Neglect
Sucheta Dalal,
16 January 2026
As the financial capital of Mumbai prepares for its upcoming municipal elections, the political discourse in Maharashtra is increasingly defined by a dual narrative. On the one hand, the ruling coalition is promoting a set of modern...
Who Paid for Electoral Bonds Printed after Supreme Court Verdict? RTIs Flag Unanswered Questions
Moneylife Digital Team
15 January 2026
A fresh set of replies received under the Right to Information (RTI) Act by transparency campaigner commodore (Cmde) Lokesh Batra (retd) has raised new questions over who ultimately bore the cost of printing high-value electoral bonds...
Supreme Court Warns States of Heavy Compensation for Dog-bite Cases, Questions Dog-feeders
Moneylife Digital Team
13 January 2026
The Supreme Court (SC) on Tuesday issued a stern warning to state governments over the rising number of stray dog attacks, signalling that it may direct them to pay 'heavy compensation' in cases involving dog bites, deaths or serious...
Bhopal to Indore: The Price of Life in ‘Viksit' Bharat
Sucheta Dalal,
09 January 2026
Indore has been celebrated as India's cleanest city for eight consecutive years. In December 2025, this reputation was shattered when contaminated sewage killed at least 10 residents and 270 others fell ill, a result of sheer...
